It has been suggested that peak aminoglycoside concentrations 4 to 8 times the minimum inhibitory concentration (MIC) are necessary to treat infections commonly found in critically ill patients, such as gram- negative sepsis and pneumonia. Consolidating conventional multiple daily doses into a single large daily dose has been proposed in order to take advantage of aminoglycoside concentration-dependent killing and to ensure peak aminoglycoside concentrations above the MIC of organisms. Using once- daily aminoglycoside dosing may be a way of attaining high peak concentrations and immeasurable trough concentrations without increased risk of nephrotoxicity and ototoxicity. Recently, Nicolau et al. studied once-daily aminoglycosides in 2,184 adult patients, and developed a once- daily dosing nomogram using a fixed volume of distribution (vd) of 0.3 L/kg. However, the study excluded patients with highly variable or altered aminoglycoside pharmacokinetic parameters, including critically ill patients. Therefore, this nomogram may not predict aminoglycoside doses in critically ill patients. The rationale for developing a dosing nomogram individualized to the ICU patient is to tailor aminoglycoside therapy using pharmacokinetic parameters specific for critically ill patients to optimize peak serum concentrations while avoiding measurable trough concentrations. This study will develop a once-daily aminoglycoside dosing nomogram specific for critically ill patients. An initial nomogram has been developed using mean aminoglycoside pharmacokinetic parameters of 134 dosing segments of 94 adult critically ill patients who received single or multiple courses of aminoglycoside therapy during their admission to the Medical Intensive Care Unit. The nomogram recommendations will be prospectively evaluated with aid of serial measurements of aminoglycoside plasma levels in patients above age 18 with gram-negative infections needing treatment with aminoglycoside. The study will exclude pediatric patients and patients with changing renal function. The study will be conducted at the National Institutes of Health, W.G. Magnuson Clinical Center, Medical Intensive Care Unit, and the Clinical Pharmacokinetics Research Laboratory. Due to the low patient census seen in the Clinical Center Medical Intensive Care Unit, the investigators also request the approval of George Washington University Hospital's Medical Intensive Care Unit as an additional site for study.